294 THE E\'OLUTION OF SCIENTIFIC THEORIES 



ways we have found to conceive and handle tables. But we are under 

 no obligation whatever to choose among these views. Each and every 

 one is justified if it sers^es the purpose(s) for which it was shaped. 

 We prize the "highest" ( quantum-mechanical ) member of the theo- 

 retical hierarchy for its immense scope of correlation, though it is 

 in fact a rather notably unuseful way of viewing the table ( save, per- 

 haps, were one concerned to use that as an impromptu radiation 

 shield). The "lowest" (common-sense) member of the theoretical 

 hierarchy is unquestionably the most useful way of viewing the table 

 as a table. Moreover, there is nothing at all illusory about this view. 

 Indeed, Eddington's "higher" table simple cannot impugn the "lower" 

 familiar view. For, as we saw earlier, the "higher" theory acquires 

 physical meaning only by presuming the absolute validity of the 

 "lower." And so, says Oppenheimer, we must recognize that— 



. . . ever-increasing refinements and critical revisions in the way we 

 talk about remote or small or inaccessible parts of the physical world 

 have no direct relevance to the familiar physical world of common 

 experience. 



EVOLUTION AND ERROR 



The emergence of higher forms of life in nature may be held to 

 signify the insuflBciency of older lower forms less competent to deal 

 with some of the range of terrestrial habitats. But the higher forms 

 de\'elop only from or through the lower which, if they be "errors," 

 are not for that the less essential to such development. The "higher" 

 scientific theory— often less convenient in everyday use, but power- 

 ful enough to impose its rule throughout a drastically extended range 

 of experience— develops similarly from and through older "lower" 

 theories less generally sufficient, and "errors" at least in the degree 

 they have been thought final. 



Comte, Mach, and their doctrinal descendents have supposed that 

 a science focused on the description, rather than on the explanation, 

 of experience would be a science built lastingly— a body of positive 

 knowledge enduring for the ages. But science seeks general knowl- 

 edge of the world, and can therefore never hope to possess knowl- 

 edge absolutely generally secure. To Comte's horror, Regnault finds 

 inexactness even in Boyle's law— an apparently exemplary economic 

 description of gaseous behavior, which makes no "vain pretense" to 

 be anything more than that. If we thus encounter limits to the suffi- 



